Metadata-Version: 2.3
Name: post-instruct
Version: 0.1.1
Summary: Turn embedding models into instruct models by adding an adaption layer
License: MIT
Author: Márton Kardos
Author-email: power.up1163@gmail.com
Requires-Python: >=3.10
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Requires-Dist: datasets (>=2.19.2,<3.0.0)
Requires-Dist: mteb (>=1.36.32,<2.0.0)
Requires-Dist: sentence-transformers (>=4.0.1,<5.0.0)
Requires-Dist: torch (>=2.6.0,<3.0.0)
Requires-Dist: tqdm (>=4.66.5,<5.0.0)
Description-Content-Type: text/markdown

# post-instruct

This project is oriented at getting superior performance with smaller embedding models with minimal costs by training an additional layer that adjusts embeddings according to an instruction.
This would also be beneficial to users, since most popular instruct embedding models require that you pass instructions as part of your query, which not only increases embedding time for each document, but prevents you from using the embeddings for multiple tasks or experiment with different instructions.


